Official Report 578KB pdf
We turn to agenda item 3, which is evidence as part of our ongoing work on understanding barriers to participation in local politics.
The Scottish Government recently published its report, following a demographic survey of local election candidates, and we are joined today by Scottish Government officials Arfan Iqbal, who is the principal researcher in constitution, international and migration analysis, and Maria McCann, who is the head of the elections team. I welcome our witnesses to the meeting. Before we turn to questions from members, I invite Maria to make a short opening statement.
Thank you very much. Good morning, everyone, and many thanks for the invitation to come to talk about the report “Local Government Candidates Survey 2022”. The idea for the survey arose from responses to the electoral reform consultation that took place in 2017 going into 2018. As part of the consultation, ministers met groups that represent women, people with disabilities and people who promote race equality. Events were also held with young people’s organisations, including the Scottish Youth Parliament. There was universal consensus on the lack of data relating to the profile of candidates at local government elections.
We recognise that gaining a better understanding of the demographic characteristics of electoral candidates and those who win elections is important in helping us to assess the representativeness of our candidates and elected members, and how that compares to the communities that they serve. Therefore, we worked with the Electoral Management Board for Scotland, the Electoral Commission, the Convention of Scottish Local Authorities and the Improvement Service, as well as a range of equality stakeholders, to develop a survey that collected diversity data on candidates standing at the May 2022 local government elections.
All 2,548 candidates who stood for election were invited to take part in the voluntary survey on candidate diversity characteristics. It is fair to say that the survey was promoted intensively. The Electoral Commission, the EMB and the Scottish Parliament political parties panel all encouraged candidates to complete the survey.
As you will be aware, 720 responses were received, which represents a response rate of 28.2 per cent. We were disappointed, of course. However, partners were not surprised because the response rate compared favourably with other similar surveys, and the results suggest that the concerns that we all share about underrepresentation are well founded. There appears to be notable divergence between the profile of respondents and that of the overall population with regard to sex, age, education, disability status and socioeconomic background. The evidence from a low response rate is valuable, so how much better would it be if we had a high response rate? The voluntary model consistently yields that level of responses, or lower. We are very interested in the committee’s views on how that might be addressed in the future.
Thank you very much, Maria. We now turn to questions from the committee. I will begin by asking about the overall process. Can you describe to the committee how the data was collected, the project timetable and the roles of local authorities, political parties and the Scottish Government in the data-collection process?
10:45
I will start with the timescales. The survey was launched in February 2022 and was closed in June 2022. On the design process, initial meetings took place with stakeholders over the summer of 2021 until the end of 2021, at which point a proposal document was published on the Scottish Government website. That document set out the intentions behind the survey and presented an opportunity for stakeholders and others to respond and provide feedback. Subsequent to that, some testing was done on the survey and some changes were made.
As I said, the survey was finally launched in February 2022. We extended the deadline for the survey to the end of June. Initially, it was planned that it would close at the start of June. That extension was to give candidates an extra opportunity to respond. From June to October, we undertook the analysis, and publication of the report was in November 2022.
Along the way, a number of stakeholder organisations and partners were involved, including, as Maria mentioned, the Electoral Management Board, the Electoral Commission, COSLA and the Improvement Service. As well as those groups, Inclusion Scotland, Engender, the Equal Representation Coalition and the Scottish Parliament political parties panel were all consulted and fed into the process. The Scottish Government had overall responsibility for carrying out the survey, analysing responses and publishing the report.
Thank you very much for that. You mentioned the work that was done prior to the survey going out. I would be interested to hear about the groups that were involved in designing the survey and the project more generally.
The Equality Coalition brings together a wide range of groups, so that was a very helpful forum. We took feedback and made amendments, and the questions went round all the bodies that I mentioned. There was a lot of very useful input, and things were adjusted in order to reflect the views of those with an interest.
I also would like to pick up on the discrepancy in respect of candidates under 34. It is welcome to see intersectional analysis being undertaken on that data, and it is welcome that we will continue to do that to fully understand the complex and multiple barriers to elected office that many people in Scotland face. One insight that I would like to hear more about is the significantly higher proportion of men than women under 34 who stand for election. That shows that the perceived progress in gender representation among younger people is not the case. I am keen to hear whether any further analysis was or can be done to identify causes for that discrepancy, so that the committee can progress work on addressing those barriers.
One of the bits of analysis that we did compared the age profiles and sex of candidates. Combining those two factors allowed us to identify that the discrepancy between males and females is particularly acute in the younger age range and in the older age range, but seems to be less prevalent among middle aged people. The research that we have undertaken is descriptive; it is not designed to allow us to determine the reasons why things are happening. It is to describe the situation as it stands, so we would just be speculating if we were to try to determine what may be driving that. If we wanted to explore that further, we would probably need to do qualitative work.
However, the survey represents the first stage in this process: in other words, it identifies where we might want to drill down further and undertake additional research.
Good morning, panel. I will just touch on the survey response rates. Is the survey response rate lower or higher than expected? What are the reasons why more than 70 per cent did not respond?
I was very ambitious in thinking that we would buck previous trends because we would reach out so much and encourage people so much. It turned out that that was not possible. We used all the avenues. The returning officers helped, along with all the points of contact for the candidates and the parties, so the return rate was really disappointing. We thought that we would get 40 per cent, 50 per cent or even more; I was so keen to do this and to make it happen. We do not really know the reasons why.
We have discussed the matter with partners. One of the reasons that was given was that people who are campaigning are caught up in the campaign, but probably see the piece of paper and think that they will do it. We got two letters from people saying that they definitely would not complete the survey, because they did not think that it was a worthwhile exercise. There were only two; it was not as though many people were writing in protesting or feeding back that they did not want to do it.
Our guess—it can be only a guess unless we drill down further, as Arfan said—is that people were caught up in the campaign. If they did not fill the survey in before they got elected, they could still have completed it, but I suppose that people who are not elected do not look back at paperwork and so on, which is very understandable. Obviously, the vast majority of candidates will be in that category because a lot of people were standing. We are just speculating, but I think that that could be the reason.
One might think that people would feel more compelled to fill in something from the Electoral Commission, just because of its status as a regulator, but partners said that they got the same kinds of responses, so they were not surprised at all. We have possibly gone as far as we can with the voluntary approach but, as you say, it is always more difficult to know why people do not do something than to know why they do something.
Thank you for that. Obviously, the survey response rates vary significantly, going from 13.9 per cent in Clackmannanshire Council to 71.4 per cent in Orkney Islands Council. Why is there such a difference?
We found that some of the contacts in the councils’ elections teams were focused on and were promoting the survey. When I worked in a local authority, I found that people were always mindful that there would be league tables for everything. Some teams would say, “Look, this will give us good performance for our council” and took ownership of the survey in that very particular way. Maybe others did not have time or were preoccupied with other things. Obviously, in the run-up to the elections, we were still in the pandemic. People were preparing in that context, so teams were under a lot of pressure. Where the survey was taken on locally is where you see the good response rates.
Thank you.
Good morning, panel. I was going to ask a question about people who did not respond and what you did, but you have kind of answered it by saying that you were using returning officers, the Electoral Commission and so on.
How confident are statisticians that the sample of candidates who responded is representative of the overall candidate population? Given that the report concludes that it is also not possible to draw firm conclusions regarding representativeness of candidates, how useful is the survey as an evidence source for policy makers?
It is really difficult for us to say anything definitive about how representative the response was because we do not have the demographic details of the people who did not respond, so we cannot do that comparison. Nonetheless, the report gives us an indicative sense of where there might be issues relating to things that Maria has already highlighted—age, sex and so on. We can cross-reference the details with similar surveys that have presented such findings. I say again that this is probably a first step towards having standardised data collection for that type of information, and it establishes a baseline. That is where the real value of the report is, as opposed to its saying anything definitive about the candidate group overall.
I will ask one more question about the people who did not respond. What else do you think could be done? What was the biggest factor, if there were big factors, that people said was the reason why they did not want to respond, other than the fact that they did not get elected and so did not want to be bothered?
The only thing that came back was that they were caught up in the campaign. We would not want to repeat the survey being done on a voluntary basis. We made the commitment to do it and, as I say, we gave it our absolute best shot, but I do not think that there was anything else that we could have done. We are interested in ideas about how things could go better in the future.
I want to pick up on that, Maria, because you have said a number of times that you do not want to do the survey again on a voluntarily basis. You have asked us, but have you got any thoughts? As a candidate, I would be thinking that I have to fill in papers that I then take to wherever—I cannot remember the name. There are papers that candidates must fill in, so is there something that we could provide so that, when they fill in the papers to register as a candidate, they also give the other information?
Yes. It would be difficult to make the survey compulsory without exception, and we would need to look into all the considerations, such as on data protection. Ideally, however, a survey would be one of the papers that candidates are required to fill in. The bottom line is that that would be the way in which we could get the data, but whether that would be acceptable is another question, and we would need to do a great deal of work.
I am interested in the committee’s views and in your instinctive reaction to that. Is it proportionate and appropriate to make it a compulsory paper? There are papers that must be filled in for a candidate’s nomination, but would making that a statutory requirement be going too far?
We will take that into our private session and give it consideration. It is a good question for us to reflect on.
Thank you.
Good morning. I will stick with the issue of the statistical significance or otherwise of the survey. Your report states that your findings
“have not been weighted or had confidence intervals applied”.
What, ultimately, does that mean? Does it mean that we cannot say that the survey was carried out using a fair and representative sample? Is that what you mean?
No, not at all. That is standard text that we put into the report to make it clear that the data has not been altered in any way. For certain types of surveys, weighting might be applied to ensure representativeness. We did not do that, so we were just making it clear to people who might want to access the underlying data that it had not been adjusted.
We would normally use such techniques if the response rate were higher, in order to ensure that the sample was representative, or in order to determine whether differences that we were seeing were significant and were not down to chance. However, owing to the low response rate, we thought it inappropriate to apply those tests, because that would have given the data an undue level of confidence. Therefore, we have not used them.
As I said, that text is there more for the technical audience, to make it clear that nothing has been done to the data but that, nonetheless, the findings can be taken at face value.
Does the fact that you did not apply those techniques compromise the data and the message that it gives?
No, it does not compromise it in any way.
I note from the response rate table in the report that the response rate of a lot of the bigger authorities was well below the average response rate. The response rates of Glasgow City Council, the City of Edinburgh Council, South Lanarkshire Council and North Lanarkshire Council are well below the average. Does that tell us something about the quality of the message? Does it say that the survey is not really representative of the broad scope of candidates who stood at the election?
As I mentioned in answer to the previous question, it is difficult for us to say anything about how representative it is, because we lack information about the 50 or 60 per cent, or whatever the figure is, of candidates in Glasgow who might not have responded. We could not compare them with those who responded to say anything definitive. It is a bit of an unknown, unfortunately.
Did most responses come in before the election or after it?
If I remember correctly, a lot of them came in before the election.
Did candidates who were not successful tend not to respond at all?
I do not think that we know that, do we?
I do not know that for sure, off the top of my head.
Unfortunately, we do not know that.
Okay. Thank you.
11:00
You have touched on some of these issues. The report presents the data nationally. Can it be broken down into sex, age, ethnicity and disability at local authority level? That would be helpful. Can we say from the survey results which groups are most underrepresented and which groups are overrepresented? You have touched on that, but is there anything more that you would like to say? Arfan, I will come to you first.
It had been our intention, if the response rate had been higher, to provide breakdowns at local authority level. Unfortunately, we decided that we could not do that because of the risk of disclosure. We had the data checked by statisticians who work for the Scottish Government, and we concluded that it would risk identifying people. For that reason, we opted not to provide that data at local authority level.
We noticed some groups that are over and underrepresented. In summary, of those who responded, we found evidence that there were fewer candidates among females, younger individuals, individuals with less than degree-level qualifications, individuals with limiting health conditions and individuals from lower socioeconomic groups, compared with the overall population. We saw one trend that was contrary to that, which was that there was a higher proportion of candidates among individuals who identified as lesbian, gay, bisexual or in some other way, compared with the overall population.
We cannot say for sure what that means but, intuitively, we might assume that people from less represented groups might be keener to partake in such surveys and make sure that their voices are heard. That might account for some degree of fluctuation.
Maria, do you want to add anything?
No, thank you.
The research focuses entirely on candidates. Is any follow-up work planned to survey those who were successfully elected to get a picture of how many female councillors there are and how many councillors there are, according to age bracket, education level and socioeconomic background? Is any follow-up work planned for the councillor cohort so that we can get a picture of what that looks like?
We understand that the Improvement Service always has a candidate survey, so it will seek to collect that data. Unfortunately, however, the response rate for that has been around 21 per cent or somewhere in the 20s, so we might not get more information from the elected members. We will need to wait and see. We will not be leading on that work—it was already in train.
Okay. I appreciate that.
The percentage of respondents who self-identified as coming from a working-class background was 11 per cent. That seems low, considering the figure for the general population. What classifications were used? How were those from a working-class background defined and identified?
I make it clear that individuals did not self-identify by class. We collected that information in line with the guidance that the UK Government’s Social Mobility Commission provided. The question that is used to determine an individual’s socioeconomic background is one that asks what the occupation of their main household earner was when they were aged 14. Research has shown that that is the best way to collect that information. Once we know those individuals’ occupations, we can group them into different categories so that we end up with the wider groups of professional background, intermediate background and working-class background. That is what has allowed us to determine respondents’ background.
The results suggested that there were more people from more affluent backgrounds among candidates than there were in the overall population. There is quite a lot of background information related to that. I would be happy to provide that in writing, if that would be of use.
That would be helpful—thank you.
Good morning. Thank you for joining us.
Further to those questions, are you satisfied that the questionnaire and the survey report meet the stated goal of considering the issue of intersectionality in relation to the profile of candidates? Has it been just as difficult to extrapolate data on that?
Yes, we had a stated goal of looking at intersectionality. If the response rate had been higher, we probably would have added more categories into that analysis. We considered age and sex, as well as sex and disability. We were able to do that because there was a sufficient number of respondents to those questions to ensure that the analysis would not have been disclosive in any way. However, given the limits due to the small response rate, we were not able to go on to look at other things, such as ethnicity and age and sex and so on. The more categories you add, the smaller the outputs would be.
We did some intersectional analysis. In the future, we would prefer to do more, but that would rely on our getting a higher response rate.
Thank you for that. The conclusion is that it will be difficult to read things into the data that we have. Has any work been done on how other countries monitor the diversity of local politicians and candidates? Has there been any learning around that?
We have looked at international comparisons. This is our first attempt, so we will look at that again. That informed our thinking at the beginning, but we need to go back and see what lessons we can learn from other countries.
That concludes our questions. Is there anything else that you want to make sure that we hear?
No, thank you.
I thank Maria and Arfan for joining us and giving evidence.
As we agreed at the start of the meeting to take the next items in private, we have no more public business. I therefore close the public part of the meeting.
11:06 Meeting continued in private until 11:42.